algo.qi dimension smaller than num_items (without CV) - Githubissues

NicolasHug / Surprise

A Python scikit for building and analyzing recommender systems

http://surpriselib.com

BSD 3-Clause "New" or "Revised" License

6.34k stars 1.01k forks source link

algo.qi dimension smaller than num_items (without CV) #390

Closed lbialik closed 3 years ago

lbialik commented 3 years ago

Description

Unlike the similar issue https://github.com/NicolasHug/Surprise/issues/294, I am not cross validating so this should not be due to downsampling.

Steps/Code to Reproduce

original matrix dim: (62926, 8324)

train_data = Dataset.load_from_df(masked_df, reader)
trainset = train_data.build_full_trainset()
algo=SVD()
algo.fit(trainset)

algo.pu.shape = (62926, 100) algo.qi.T.shape = (100, 8245)

Versions

import platform; print(platform.platform()) Linux-3.10.0-1127.19.1.el7.x86_64-x86_64-with-glibc2.10 import sys; print("Python", sys.version) Python 3.8.8 (default, Feb 24 2021, 21:46:12) [GCC 7.3.0] import surprise; print("surprise", surprise.version) surprise 1.1.1

lbialik commented 3 years ago

Fixed: it was dropping columns with no non-zero values.